[GRASS5] pbmplus headers for r.in/out.png

Eric G . Miller egm2 at jps.net
Fri Jan 26 11:00:57 EST 2001

On Fri, Jan 26, 2001 at 04:17:19PM +0100, Andreas Lange wrote:
> Hi,
> i checked in the netpbm/pbmplus header files to src/include.
> Could someone please check if this is ok with the copyright in the
> netpbm distribution, i am not sure. Please remove if not.
> r.in.png and r.out.png compile on linux and IRIX, but i could not test
> both modules. The code is a bit confusing (mildly put). 
> NB: r.in.tiff still segfaults on my linux machine (RH 7.0, Intel
> processor). Could not test on IRIX.

   I had started a complete rewrite of r.in.tiff.  However, I got stuck
on handling the TIFF world file a little. If the X-shift/Y-shift
parameters are not zero, then an Affine Transformation needs to be
applied.  I got stuck at figuring out what the cell resolution of the
resulting CELL file is supposed to be (I only know it should come out
square).  I searched and searched, but could not find an example of how
to calculate this crucial bit of information given the coefficients to
the Affine Transformation Formulas:

     x' = Ax + By + C, y' = Dx + Ey + F

     where the 'By' and 'Dx' terms affect rotation and shearing.

I know there's a relatively simple solution somewhere.  I just never
found it.  Hence, I've put r.in.tiff on the back burner.  I know the
solution already exists in the GRASS code base, since i.rectify does
this very task, I just wanted to be able to handle it directly rather
than having to pass off the task to another module...

BTW: r.in.tiff mostly works here.  It's really fragile though.  It tends
to make the bad assumption of using one byte per band rather than an
unsigned short (which is allowed in the baseline 6.0 spec.).  This
doesn't seem to be a problem for palette based rasters, but it
definitely can be a problem for true color.

Eric G. Miller <egm2 at jps.net>

If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo at geog.uni-hannover.de with
subject 'unsubscribe grass5'

More information about the grass-dev mailing list